Hello Krzysztof,
On 03/01/2016 08:44 PM, Krzysztof Kozlowski wrote:
On 02.03.2016 01:45, Javier Martinez Canillas wrote:
The watchdog maximum timeout value is determined by the number of bits
for the interval timer counter, its source clock frequency, the number
of bits of the prescaler and maximum divider value.
This can be calculated with the following equation:
max_timeout = counter / (freq / (max_prescale + 1) / max_divider)
Setting a maximum timeout value will allow the watchdog core to refuse
user-space calls to the WDIOC_SETTIMEOUT ioctl that sets not supported
timeout values.
For example, systemd tries to set a timeout of 10 minutes on reboot to
ensure that the machine will be rebooted even if a reboot failed. This
leads to the following error message on an Exynos5422 Odroid XU4 board:
[ 147.986045] s3c2410-wdt 101d0000.watchdog: timeout 600 too big
Reported-by: Krzysztof Kozlowski <k.kozlowski@xxxxxxxxxxx>
Signed-off-by: Javier Martinez Canillas <javier@xxxxxxxxxxxxxxx>
---
Hello,
I'm sending this as an RFC because I don't have access to user manuals
for all the Exynos SoCs that this driver supports. But at least in all
the ones I have here, the maximum prescaler and divider values are the
same and also all have a 16-bit interval timer counter.
So this patch assumes that all IP blocks supported by this driver are
the same on that regard, if that's not the case then we can move these
values on a per IP block basis using struct of_device_id .data field.
It looks correct at least for Exynos4210 and newer.
Great, thanks a lot for the confirmation.
Best regards,
Javier
drivers/watchdog/s3c2410_wdt.c | 16 ++++++++++++++++
1 file changed, 16 insertions(+)
diff --git a/drivers/watchdog/s3c2410_wdt.c b/drivers/watchdog/s3c2410_wdt.c
index 0093450441fe..f289c9fc353a 100644
--- a/drivers/watchdog/s3c2410_wdt.c
+++ b/drivers/watchdog/s3c2410_wdt.c
@@ -47,6 +47,8 @@
#define S3C2410_WTDAT 0x04
#define S3C2410_WTCNT 0x08
+#define S3C2410_WTCNT_MAXCNT 0xffff
+
#define S3C2410_WTCON_RSTEN (1 << 0)
#define S3C2410_WTCON_INTEN (1 << 2)
#define S3C2410_WTCON_ENABLE (1 << 5)
@@ -56,8 +58,11 @@
#define S3C2410_WTCON_DIV64 (2 << 3)
#define S3C2410_WTCON_DIV128 (3 << 3)
+#define S3C2410_WTCON_MAXDIV 0x80
+
#define S3C2410_WTCON_PRESCALE(x) ((x) << 8)
#define S3C2410_WTCON_PRESCALE_MASK (0xff << 8)
+#define S3C2410_WTCON_PRESCALE_MAX 0xff
#define CONFIG_S3C2410_WATCHDOG_ATBOOT (0)
#define CONFIG_S3C2410_WATCHDOG_DEFAULT_TIME (15)
@@ -198,6 +203,14 @@ do { \
/* functions */
+static inline unsigned int s3c2410wdt_max_timeout(struct clk *clock)
+{
+ unsigned long freq = clk_get_rate(clock);
+
+ return S3C2410_WTCNT_MAXCNT / (freq / (S3C2410_WTCON_PRESCALE_MAX + 1)
+ / S3C2410_WTCON_MAXDIV);
+}
+
static inline struct s3c2410_wdt *freq_to_wdt(struct notifier_block *nb)
{
return container_of(nb, struct s3c2410_wdt, freq_transition);
@@ -567,6 +580,9 @@ static int s3c2410wdt_probe(struct platform_device *pdev)
return ret;
}
+ wdt->wdt_device.min_timeout = 1;
+ wdt->wdt_device.max_timeout = s3c2410wdt_max_timeout(wdt->clock);
Can the frequency of clock change? E.g. with devfreq? No problem if it
goes lower but if it gets higher than initial, then the problem will
appear again.
That's a very good question. As Guenter said we will be in deep troubles
if that ever happens since the driver doesn't take that into account.
The .set_timeout handler just sets the counter according to the current
frequency and that's never updated, unless a new timeout is set of course.
So in other words, I just made the same assumptions that the driver is
currently doing. At least the Exynos SoCs manual don't mention frequency
scaling for the watchdog timer source clock and AFAICT none of the CLK_WDT
parents scale their frequencies but I don't know if that's true for all
the machines using this driver (i.e: out-of-tree boards).
Best regards,
Krzysztof
Best regards,
--
Javier Martinez Canillas
Open Source Group
Samsung Research America
--
To unsubscribe from this list: send the line "unsubscribe linux-samsung-soc" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
